Feature Extraction 主講人:虞台文
Content Principal Component Analysis (PCA) Factor Analysis Fisher’s Linear Discriminant Analysis Multiple Discriminant Analysis
Principal Component Analysis (PCA) Feature Extraction Principal Component Analysis (PCA)
Principle Component Analysis It is a linear procedure to find the direction in input space where most of the energy of the input lies. Feature Extraction Dimension Reduction It is also called the (discrete) Karhunen-Loève transform, or the Hotelling transform.
The Basis Concept x w wTx That is, Demo Assume data x (random vector) has zero mean. w PCA finds a unit vector w to reflect the largest amount of variance of the data. wTx That is, Demo
The Method Remark: C is symmetric and semipositive definite. Covariance Matrix
The Method maximize subject to The method of Lagrange multiplier: Define The extreme point, say, w* satisfies
The Method maximize subject to Setting
Discussion At extreme points Let w1, w2, …, wd be the eigenvectors of C whose corresponding eigenvalues are 1≧ 2 ≧ … ≧ d. They are called the principal components of C. Their significance can be ordered according to their eigenvalues. w is a eigenvector of C, and is its corresponding eigenvalue.
Discussion At extreme points Let w1, w2, …, wd be the eigenvectors of C whose corresponding eigenvalues are 1≧ 2 ≧ … ≧ d. They are called the principal components of C. Their significance can be ordered according to their eigenvalues. If C is symmetric and semipositive definite, all their eigenvectors are orthogonal. They, hence, form a basis of the feature space. For dimensionality reduction, only choose few of them.
Applications Image Processing Signal Processing Compression Feature Extraction Pattern Recognition
Example Projecting the data onto the most significant axis will facilitate classification. This also achieves dimensionality reduction.
Issues The most significant component obtained using PCA. PCA is effective for identifying the multivariate signal distribution. Hence, it is good for signal reconstruction. But, it may be inappropriate for pattern classification. The most significant component for classification
Whitening Whitening is a process that transforms the random vector, say, x = (x1, x2 , …, xn )T (assumed it is zero mean) to, say, z = (z1, z2 , …, zn )T with zero mean and unit variance. z is said to be white or sphered. This implies that all of its elements are uncorrelated. However, this doesn’t implies its elements are independent.
Whitening Transform Decompose Cx as Set Clearly, D is a diagonal matrix and E is an orthonormal matrix. Whitening Transform Let V be a whitening transform, then Decompose Cx as Set
Whitening Transform Proof) If V is a whitening transform, and U is any orthonormal matrix, show that UV, i.e., rotation, is also a whitening transform. Proof)
Why Whitening? With PCA, we usually choose several major eigenvectors as the basis for representation. This basis is efficient for reconstruction, but may be inappropriate for other applications, e.g., classification. By whitening, we can rotate the basis to get more interesting features.
Feature Extraction Factor Analysis
What is a Factor? If several variables correlate highly, they might measure aspects of a common underlying dimension. These dimensions are called factors. Factors are classification axis along which the measures can be plotted. The greater the loading of variables on a factor, the more that factor can explain intercorrelations between those variables.
Graph Representation Quantitative Skill (F1) Verbal (F2) 1 +1
What is Factor Analysis? A method for investigating whether a number of variables of interest Y1, Y2, …, Yn, are linearly related to a smaller number of unobservable factors F1, F2, …, Fm. For data reduction and summarization. Statistical approach to analyze interrelationships among the large number of variables & to explain these variables in term of their common underlying dimensions (factors).
Example What factors influence students’ grades? Observable Data Quantitative skill? unobservable Example Verbal skill? Observable Data
The Model y: Observation Vector B: Factor-Loading Matrix f: Factor Vector : Gaussian-Noise Matrix
The Model y: Observation Vector B: Factor-Loading Matrix f: Factor Vector : Gaussian-Noise Matrix
The Model Can be obtained from the model Can be estimated from data
The Model Commuality Specific Variance Explained Unexplained
Example Cy BBT + Q =
Goal Our goal is to minimize Hence,
Uniqueness Is the solution unique? There are infinite number of solutions. Since if B* is a solution and T is an orthonormal transformation (rotation), then BT is also a solution.
Cy = Example Which one is better?
Example Left: each factor have nonzero loading for all variables. Right: each factor controls different variables. i1 i2 i1 i2
The Method Determine the first set of loadings using principal component method.
Example Cy
Factor Rotation Factor-Loading Matrix Rotation Matrix Factor Rotation:
Factor Rotation Criteria: Varimax Quartimax Equimax Orthomax Oblimin Factor-Loading Matrix Factor Rotation:
Criterion: Maxmize Varimax Subject to Let . . .
Criterion: Maxmize Varimax Subject to Construct the Lagrangian
Varimax cjk dk bjk
Varimax Define is the kth column of
Varimax is the kth column of
Varimax Goal: reaches maximum once
Varimax Goal: Initially, obtain B0 by whatever method, e.g., PCA. set T0 as the approximation rotation matrix, e.g., T0=I. Iteratively execute the following procedure: evaluate and You need information of B1. find such that Next slide if stop Repeat
Varimax Goal: Pre-multiplying each side by its transpose. Initially, obtain B0 by whatever method, e.g., PCA. set T0 as the approximation rotation matrix, e.g., T0=I. Iteratively execute the following procedure: evaluate and You need information of B1. find such that Next slide if stop Repeat
Varimax Criterion: Maximize . . .
Maximize Varimax Let
Fisher’s Linear Discriminant Analysis Feature Extraction Fisher’s Linear Discriminant Analysis
Main Concept PCA seeks directions that are efficient for representation. Discriminant analysis seeks directions that are efficient for discrimination.
Classification Efficiencies on Projections
Criterion Two-Category 1 m 2 m
Scatter ||w|| = 1 w m m The larger the better Between-Class Scatter Between-Class Scatter Matrix Scatter ||w|| = 1 w 1 m Between-Class Scatter 2 m The larger the better
Scatter ||w|| = 1 w m m The smaller the better Within-Class Scatter Between-Class Scatter Matrix Scatter Within-Class Scatter Matrix ||w|| = 1 w 1 m 2 m Within-Class Scatter The smaller the better
Goal ||w|| = 1 w m m Define Generalized Rayleigh quotient Between-Class Scatter Matrix Goal Within-Class Scatter Matrix ||w|| = 1 Define Generalized Rayleigh quotient w 1 m 2 m The length of w is immaterial.
Generalized Eigenvector To maximize J(w), w is the generalized eigenvector associated with largest generalized eigenvalue. Define Generalized Rayleigh quotient That is, or The length of w is immaterial.
Proof To maximize J(w), w is the generalized eigenvector associated with largest generalized eigenvalue. Set That is, or
Example 2 1 m - w w w
Multiple Discriminant Analysis Feature Extraction Multiple Discriminant Analysis
Generalization of Fisher’s Linear Discriminant For the c-class problem, we seek a (c1)-dimension projection for efficient discrimination.
Scatter Matrices Feature Space Total Scatter Matrix 2 m 1 m + Within-Class Scatter Matrix 3 m Between-Class Scatter Matrix
The (c1)-Dim Projection The projection space will be described using a d(c1) matrix W. 2 m 1 m + 3 m
Scatter Matrices Projection Space Total Scatter Matrix 2 m 1 m 1 ~ m 2 3 + + Within-Class Scatter Matrix 3 m W Between-Class Scatter Matrix
Criterion Total Scatter Matrix Within-Class Scatter Matrix Between-Class Scatter Matrix